even when the weather is not stormy, and in most of the large lakes minor fluctuations occur, the level of the lake rising and falling a few inches every ten, twenty or thirty minutes, as the case may be.

During the summer when the lakes become thermally stratified with a stratum of cold and relatively heavy water at the bottom and warm lighter water at the top, the seiches produced by the wind in the upper strata gradually transfer their motion to the heavier water beneath, so that this also acquires an oscillatory motion, the amplitude and period of which may be different from that of the surface seiche. These variable fluctuations tend to mix the water of the two strata in the region of the thermocline.

The seiches have an important influence on the undertow currents described beyond.

Seiche movements may occur both longitudinally and transversely at the same time, in fact, this is probably the rule rather than the exception. The period of the transverse seiche is shorter than the other, so that along the shore in the middle of the lake the slopes of the water surface take various directions according to the manner in which the crests of the seiche movements coincide In consequence of this diagonal currents may be set up along the shore, which change frequently in force and direction, and which apparently bear but little relation to the direction of the wind blowing at the time. This is of particular importance in explaining the nature of the currents that occur after the wind which produced the seiche has ceased.

VERTICAL CURRENTS

The water of the Great Lakes show, in general, the same fluctuations in the temperature of the water at the surface and at depths below the surface, as have been so frequently observed in smaller lakes and reservoirs. That is, during the summer the surface water is warmed by the sun, while the water at the depths is cold and quiescent. The wind stirs up the water of the lake and keeps it thoroughly mixed for a number of feet down from the surface. In small bodies of water the depth to which this wind action extends is seldom more than twenty or thirty feet, but in the large, deep lakes it is more. Observations made in Lake Erie at Cleveland, in 1904, showed that the lake water was stirred by the summer winds to a depth of about fifty feet, while observations made in

1

י

Lake Michigan in 1910 showed that the water was stirred to a depth of about sixty feet. Below the stratum within which circulation takes place, the temperature of the water changes rapidly and the region where this change occurs is known as the thermocline. Below the thermocline the bottom water in the middle of the lakes probably remains relatively quiescent for long periods, but for several miles out from the shore the lower strata move in and out and are seldom at rest.

The temperature of the water at the bottom of the Great Lakes during the period of summer stagnation is practically that of maximum density, i. e., 39.2° F. The author once measured the temperature at the bottom of Lake Champlain (396 feet) in June and found it to be 39.3°. Comparatively few observations of the temperature of the water at the bottom of the Great Lakes have been made, but they substantiate this statement. Recently observations in Lake Ontario near Rochester have shown that water of maximum density came within a mile and a half of the shore at a depth of forty feet.

During the winter the surface water is colder than the bottom water. The latter maintains its temperature at about maximum density throughout the year.

During the spring, as well as during the fall, there is a period when the water has an opportunity to circulate from top to bottom. In very large lakes this period is short, and whether circulation actually extends to the bottom is not known.

Besides the vertical circulation produced by the natural heating and cooling of the surface water and the action of the wind vertical circulation may also take place near the shore, due to the inflow of river water or sewage of a different temperature from the water of the lake. This sometimes has an important effect upon the manner in which these inflowing streams mix with the lake water.

Thermal stratification of the water in small lakes has an important effect upon the depth of the return currents above referred to, causing the back flow to take place above the layer of stagnant water. But in the Great Lakes the thermocline marks the dividing line between the upper and lower currents moving in opposite directions. It is an unstable thermocline and does not necessarily imply stagnation.

mike from the snore during the ten-hour period, a distance The rare water woud, therefore, be the period was 3% of the tuan eulbornt to carry it to the longest intake now in use. In widson to the velocity of the surface water induced by the wond, it is important to know al-o the angle of dispersion. Aa attempt was made to measure this at Cleveland, in 1905. of foute was wet, adrift in Lake Erie, at a time when the wind was blowing at the rate of thirteen miles per hour, and the position of A group ench font observed for several hours. In one experiment after the fonts had traveled 2,500 feet, which was 4.7% of the wind move

t, they had spread through an angle of about 5°. In other riments with higher wind velocities, the angle of dispersion was atly less. At the present time no data exist showing the relabetween wind velocity and the angle of dispersion. It varies ely according to the veering of the wind and may be as high as or 60° more. After several days of variable winds the fan of luted water may spread over a semi-circle.

UNDERTOW CURRENTS.

When the wind blows strongly towards a lee shore the raised ater level that results causes an outward current of the water neath the surface. This phenomenon has long been familiar to shermen and bathers. It has not been fully realized, however, hat the undertow currents sometimes extend for several miles into he lake and well beyond the ordinary water supply intakes. As a esult shore water is carried outward and sewage contaminated vater periodically enters the intake, and is pumped into the city. nains. The best evidence of these undertow currents is obtained by a study of the temperature of the water at different depths, at different distances from the shore and of the temperature of the water drawn from the intakes.

Daily observations of the temperature of the water at Milwaukee and elsewhere during the summer have shown very great changes in the course of a few days. Thus at Milwaukee, the temperature of the water at the pumping was 68° on July 20, 1910; on July 23, it was 62°; and on July 27th, 52°. At Rochester, even more sudden changes have been noticed. For example, on August 8, 1912, the temperature of the water at the bottom of the lake at a point 6,000 feet from the shore, where the depth was forty feet, the temperature fell from 59° to 42° in eight hours, the surface temperature meanwhile remaining about 64°. Two days previous, at the same distance from the shore, the temperature of the water had been 63° at all depths. The drop in temperature followed a sudden change in the direction of the wind from on-shore to off-shore.

The data now at hand indicate that the dividing line between the incoming and outgoing water lies at about mid depth.

When the warm surface water is blown away from the shore and cold water flows landward near the bottom, not all of the bottom water flows in as far as the shore line. Some of it becomes mixed with the upper strata and returns to the lake before it reaches the

LOCAL WIND CURRENTS.

From the standpoint of the pollution of water supplies, the most important currents that take place in the lakes are the local currents produced by the movement of the wind from day to day. As the wind blows over the lake surface currents of the water are induced by it, and as the wind changes, these currents also change. Observations that have been made in Lake Erie, notably by Walter P. Rice, at Cleveland, in 1890, and by the Cleveland Water Department, in 1895, indicate that there is a fairly close relation between the wind movement and the movement of the surface water; the travel of the surface water being commonly from 4 to 6% of the wind movement and averaging about 5%. The movement of the water is less at depths below the surface, decreasing with the depth and presumably becoming zero at the neutral line, which recent studies in Lake Ontario have shown to be at mid-depth near the shore. Observations made in Owasco Lake, near Auburn, N. Y., in 1911, by Mr. J. Walter Ackerman and the author, showed that the percentage which the travel of the surface water was of the wind movement varied with the velocity of the wind, being 3.2% when the wind velocity was five miles per hour, but only 1.2% when the velocity was thirty miles per hour. These observations were, however, made in a lake of relatively small size.

The velocities induced by the wind are greater on the lee shore than on the weather shore.

In order to illustrate the effect of these local currents induced by the wind, let it be assumed that the wind has been blowing off shore for ten hours at the rate of twenty miles per hour in such a direction as to cause the water to flow from the sewer outfall towards the water supply intake, and let it be assumed further that the velocity of the surface water during this period was 3% of the wind movement. The surface water would, therefore, be carried six miles from the shore during the ten-hour period, a distance more than sufficient to carry it to the longest intake now in use.

In addition to the velocity of the surface water induced by the wind, it is important to know also the angle of dispersion. An attempt was made to measure this at Cleveland, in 1905. A group of floats was set adrift in Lake Erie, at a time when the wind was blowing at the rate of thirteen miles per hour, and the position of each float observed for several hours. In one experiment after the floats had traveled 2,500 feet, which was 4.7% of the wind move